top of page

Ad Testing: Facial Recognition of Emotion

Project Overview

 

The Video Ad Emotion Analysis Project, conducted in collaboration with Realeyes in August 2017, aimed to assess the emotional responses of viewers to four video ads. This strategic project involved testing and refining creative content to optimize media planning and ensure effective targeting and budget allocation.

 

Objectives

 

The primary objectives of the project were to:

  1. Evaluate Emotional Responses: Measure and analyze the emotional reactions elicited by four video ads to understand their effectiveness.

  2. Benchmark Performance: Compare the emotional scores of the ads against global norms to identify strengths and areas for improvement.

  3. Optimize Media Planning: Use insights from the emotional analysis to inform strategic media spending and targeting decisions.

 

Implementation

 

  • Research Partner: Realeyes - an emotion testing AI solution provider.

  • Test Criteria: Four videos were tested in two pairs, with each video viewed by 300 participants in the USA. The participants were men and women aged 18-69.

  • EmotionAll® Scores: The videos were evaluated based on four key performance indicators (KPIs): Attraction, Retention, Engagement, and Impact. The scores ranged from 1 to 10, with benchmarking against a global database of approximately 10,000 videos.

  • Data Collection: Emotional responses were recorded and analyzed, focusing on positive and negative sentiments, engagement levels, and demographic preferences.

 

Key Results

 

  • Performance Summary:

    • Study 1: 'Primadonna' and 'Teenage Llama' performed above average, with high scores in Retention, Engagement, and Impact.

    • Study 2: 'Comma' performed at average levels, while 'Diorama' scored below average.

  • Audience Engagement:

    • Demographics: Higher engagement was observed among the 50-69 age group for Study 1 videos, and the 30-49 age group for Study 2 videos. Women showed higher engagement overall, except for 'Comma'.

    • Emotional Peaks: Humorous elements in the videos led to significant emotional peaks, particularly at punchline moments.

  • Benchmark Performance: 'Primadonna' and 'Teenage Llama' exceeded normative levels in happiness and engagement metrics. 'Diorama' had the highest negative sentiment scores.

​

See example outputs here.

 

Outcomes and Impact

 

  • Enhanced Ad Effectiveness: Insights from the emotional analysis highlighted the strengths and weaknesses of each video, enabling the retailer to refine their creative content.

  • Targeted Media Spend: The findings guided strategic decisions to prioritize media spend on the most effective videos ('Primadonna' and 'Teenage Llama') and target specific age groups more effectively.

  • Improved Customer Engagement: Understanding the emotional drivers behind engagement helped Zappos tailor their messaging to better resonate with their audience.

 
Conclusion

 

The Video Ad Emotion Analysis Project successfully provided actionable insights into the emotional impact of the retailer's video advertisements. By leveraging these insights, the retailer enhanced the effectiveness of their marketing campaigns, optimize media spending, and engage their audience more meaningfully. This project underscores the value of emotional analysis in crafting strategic business decisions and improving customer experience.

​

See full report.

bottom of page